Mean–variance optimization of a single portfolio, as introduced by Markowitz (1952, 1959), is well studied and well understood. Its influence can be found in many branches of the quantitative finance community (Fabozzi et al. 2002). Advances in mathematical programming techniques have not only allowed for the fast and reliable solution of large-scale mean–variance optimization problems, but have also allowed for the incorporation of many relevant business considerations, such as limits on the number of names, threshold position levels, and even long/short ratios. As a result, several commercial software vendors now provide asset management solutions based on mean–variance optimization.